Research highlights 1999

(for Numbered publications see Appendix 2)

Carbonatites and kimberlites ? melt inclusions from the deep lithosphere

GEMOC'S  ONGOING COLLABORATION with Kennecott Canada Inc. has shown that the lithosphere beneath the Lac de Gras area of the Slave Craton (northern Canada) consists of two distinct layers (figure 1, Publication 137).  Among the xenoliths brought up by the kimberlites in this area there are many cm- to dm-sized megacrysts of chrome-diopside pyroxene (figure 2). Mineral inclusions in the pyroxenes show that they are fragments of very coarse-grained garnet peridotites derived from the base of the lithosphere (180-200 km depth).  Many of these contain abundant inclusions of quenched carbonatite melts, as well as inclusions of ultramafic silicate melts and Fe-Ni sulfide melts. These melts were trapped as fluid inclusions in the clinopyroxene, which shielded them from interaction with kimberlite and mantle and crustal wall-rocks during ascent. The microstructures of the carbonatites, in particular, show that they were quenched in near-surface conditions.  The relationships between the carbonatitic, ultramafic silicate and sulfide melts are complex, with all three types occurring along apparent single inclusion trails through the pyroxenes.  At present, the data suggest that all of the inclusions were derived from a carbonated ultramafic parent melt by progressive degrees of liquid immiscibility, while cooling through a small temperature range near 1250 °C.

Figure 1.  Chemical tomography images of the lithospheric mantle beneath the Lac de Gras area of the Slave Craton.  Top: Zr, Y and Ti contents of garnets show  the boundary between the ultradepleted upper layer of the lithosphere and the less depleted deeper layer.  Bottom:  Rock types deduced from garnet chemistry show that depleted harzburgites are restricted mostly to the shallow layer, while lherzolites dominate the lower layer.
GEMOC's work on these unique samples has integrated data from several components of the Facility for Integrated Microanalysis (see Technology Development).  The electron microprobe and the scanning nuclear microprobe have been used to image the structure and the distribution of major and trace elements, and the LAM-ICPMS has been used for detailed trace-element analysis at low levels.  Sr isotope ratios will be measured using the LAM-MC-ICPMS.

The carbonatite inclusions show radiating "spinifex" olivine and blocky phlogopite crystals, set in a matrix of calcite; two generations of calcite with different Sr and Ba contents are present, reflecting the concentration of these elements in the first carbonate to crystallise.  Detailed image analysis of modal compositions, combined with analysis of the individual phases, shows the bulk composition of the melts to be a magnesian silico-carbonatite.

In some ultramafic inclusions, the silicate and carbonate components have unmixed into concentric rings in a single globule (figure 2), where the carbonatite separates an outer ring of ultramafic melt (UM) from a more magnesian silicate melt in the core.  Many of these inclusions contain abundant dispersed Fe-Ni sulfide.  The scanning nuclear microprobe images show the partitioning of different elements among these components.  The calculated bulk compositions of these inclusions closely resembles that of many kimberlites.  The study of these unusual samples is providing new insights into the processes that produce both kimberlites and carbonatites, and on their metasomatic effects in the mantle.

Contact:  Esmé van Achterbergh, Bill Griffin,
Sue O'Reilly, Norm Pearson
Funded by:  ARC, Kennecott Canada Inc. and Macquarie University.
Part of GEMOC's Lithosphere Mapping, Geotectonics and Metallogenesis strands.

Figure 2.  Centre:  Carbonatite inclusions in a chrome-diopside megacryst (1.5 x actual size).  Left:  Backscattered  electron (BSE) image and maps of Sr and Ni distribution in a carbonatite bleb, produced with the CSIRO-GEMOC Scanning Nuclear Microprobe.  Right:  BSE image of composite carbonatite-kimberlite blebs, and maps of Mg and Ca distribution produced by energy-dispersive electron microprobe analysis.

Sulfides - a shining tracer of lithosphere evolution.

THE PLATINUM GROUP ELEMENTS  (Os, Ir, Ru, Rh, Pt, Pd), together with Au and Re, comprise the Highly Siderophile Elements (HSE). The relative abundances of these elements are important for models of the formation and evolution of the Earth. Because of their siderophile and variably chalcophile nature they also provide, along with the Re-Os isotopic system, a different perspective on the formation and evolution of the different parts of the Earth (Core, Mantle and Crust). Sulfide minerals are ubiquitous in the mantle and, despite their very low modal abundance (<0.1%), have been inferred to be the main host phases for the HSE. However, the behaviour and trace element contents of these sulfides is very poorly known.

A detailed study of several mantle xenolith suites world wide (in collaboration with Dr. J.P. Lorand, Muséum National d'Histoire Naturelle de Paris), has shown that sulfides are sensitive to many processes, including melt removal and metasomatism (figure 1). Heating events and extensive melt/rock reaction trigger sulfur loss because of the low melting temperature of sulfide, while volatile-rich metasomatism may add significant amounts of sulfide. All these features suggest that the HSEs, hosted in sulfides, can be significantly mobile in the mantle.

Figure 1. Effects of different processes on mantle sulfides

The in situ analytical capabilities of the GEMOC Facility for Integrated Microanalysis (Laser ablation microprobe (LAM)-ICPMS, LAM-MC-ICPMS and Scanning Nuclear Microprobe) have allowed us to make a breakthrough in the understanding of the behaviour of the HSEs. In situ analyses of enclosed and interstitial sulfides yield contrasting HSE patterns. Silicate-enclosed sulfides, shielded from fluid percolation processes, have mono-sulfide solid solution compositions, high HSE abundances and Pd/Ir ratio less than chondrites (figure 2). In contrast, interstitial sulfides have a Cu-rich pentlandite composition, lower HSE contents and high Pd/Ir ratios (figure 2). Both sulfides often occur in the same sample. Our data indicate that silicate-enclosed sulfides are the residua of melting process(es), while interstitial sulfides are the result of crystallisation of sulfide-bearing fluids. Hence, we suggest that non-chondritic HSE patterns are due to process(es) occurring in the upper mantle (i.e. melting and sulfide addition via metasomatism), and are not evidence of core material addition or Űexotic' meteoric bombardments as has been recently proposed.

The results are also significant for the interpretation of Re-Os depletion ages on mantle xenoliths, because the mobility of sulfide implies the mobility of Os, and the mixing of material from different sources will produce ambiguous age data.  Preliminary in situ Re-Os dating of mantle sulfides (see News Flash) shows that enclosed and interstitial sulfides may have widely different Os isotope compositions, and that whole rock 187Os/188Os may reflect mixing between several sulfide populations rather than "young" melt depletion ages. A detailed knowledge of the petrography and chemistry of the sulfide component of the rock has become essential to understanding and interpreting Re-Os data.

Contact:  Olivier Alard, Simon Jackson, Bill Griffin, Sue O'Reilly.
Funded by:  ARC and Macquarie University
Part of GEMOC's Lithosphere Mapping and Metallogenesis strands.

Figure 2. Sulfide inclusions in mantle silicate have PGE patterns showing Os enrichment and may date mantle depletion events.  Interstitial sulfides in mantle rocks show PGE patterns that are flat to Pd-enriched and track metasomatic effects.

The Re-Os isotopic system has rapidly become an important tool for dating the timing of melt-extraction events in the upper mantle.  The recognition and definition of two types of Platinum Group Element (PGE) patterns in mantle sulfides (see elsewhere in Research Highlights) has important implications for the interpretation of Re-Os ages on mantle-derived xenoliths.  That work also has shown that many grains of Fe-Ni sulfides enclosed in the primary minerals of mantle xenoliths have very high Os contents.

Late in 1999, we demonstrated that these grains could be dated by in situ laser-ablation analysis on the GEMOC MC-ICPMS.  A series of test analyses on a synthetic Fe-Ni sulfide standard spiked with PGEs and with Os of known isotopic composition showed that these analyses could be done in a few minutes, with precision and accuracy comparable to standard TIMS analyses.  The technique was then applied to large (200 micron) sulfide grains included in olivine xenocrysts from the Udachnaya kimberlite, Siberia (via Prof. Y. Barashkov) and the A481 kimberlite, Slave Province (via Jennifer Burgess of Diavik Diamond Mines).  These gave similar model ages (TMA) for melt removal:  3.07±0.05 Ga (2sd) for Udachnaya and 3.01±0.07 Ga  for A481. Both are equivalent to the oldest ages reported for xenoliths from these areas by conventional TIMS Re-Os analysis. Preliminary analyses of interstitial  sulfides in several types of mantle xenoliths show that these have much more radiogenic Os than the enclosed sulfides.  Conventional whole-rock analyses will represent mixtures of these types, with ambiguous ages.

The LAM-MC-ICPMS technique is very rapid and inexpensive, compared to conventional TIMS analysis, so that dating of mantle sections can be done on a statistical basis.  Even more important, the spatial resolution of the in situ technique means that the data can be interpreted with less ambiguity, since a single generation of sulfide is being analysed with each shot.  The new technique also will make dating of single diamonds a relatively simple matter - and sulfides are the most common inclusion in diamonds!

Contact:  Bill Griffin, Norm Pearson, Olivier Alard, Sue O'Reilly Funded by:  ARC
Part of GEMOC's Lithosphere Mapping and Metallogenesis strands.

Composition of groundwaters: geochemical fingerprints of aquifer lithology

TRANSPORTED SOIL presents an exploration problem in many parts of Australia, hiding geological outcrop and obscuring the geochemical signatures of ore deposits.  Groundwaters offer a unique opportunity to see beneath cover and interpret basement geology.  Groundwaters are active exploration media in that they can move vertically and laterally, sampling large volumes of rock by dissolving major and trace elements in amounts depending on the local geological and geothermal environment.  Until recently there has been a problem in relating aquifer rock type with groundwater major element chemistry in a simple way that is amenable to exploration.  A GIS approach with large groundwater geochemical databases is addressing this problem.

Figure 1.  Schoeller plot of groundwater geochemistry with respect to aquifer lithology.

By considering approximately 37,000 groundwaters and linking their locations with known surface geology, a series of major element concentration plots (Schoeller plots) have been made (figure 1).  Each Schoeller plot gives a distinct groundwater signature of the rock type with which it is assumed to be associated.  Both the position and shape of the line traces can be used to distinguish between signatures.  Schist, gneiss, limestone and ultramafic aquifers result in unique patterns.  The relative ratios of consecutive ions (e.g. SO42- and HCO3-) are particularly important when establishing differences between basalts and metabasalts.  No differences were expected (or found) between granitic and silicic volcanic aquifers.

Schoeller plots provide a simple, effective way of distinguishing groundwaters from different sources.  This can have exploration applications and has been demonstrated for paleochannel uranium deposits. Schoeller plots from mineralised paleochannels are significantly different from those from the surrounding area, even when close to uranium mineralisation of a different style.

Aquifer lithology dictates the most appropriate geothermometer to apply in a particular environment.  Current research is using groundwater geochemistry to suggest aquifer lithology in parts of the Great Artesian Basin (figure 2).  This information will be used to select the most appropriate geothermometer prior to making an estimation of subsurface temperature.  This, combined with bore depths and thermal conductivity measurements, will allow heat flow to be estimated across the basin.  New field data, combined with data from state government databases in NSW, Qld and SA are being used to achieve this.

Contact:  Mark C Pirlo, Sue O'Reilly
Project funded and supported by: Queen's Trust for Young Australians Achievement Award, GEMOC National Key Centre, APA award, MUPR Grant
Part of GEMOC's Lithosphere Mapping and Metallogenesis strands.

Figure 2.  Blanche Cup (foreground) and Hamilton Hill (background): natural discharge points of the Great Artesian Basin, South Australia.

Copper isotope geochemistry - new tools for exploration and research

STABLE ISOTOPES have long been used to study the sources of O, C and S in mineralising systems.  The isotopes of the commodity elements have been ignored because of the analytical limitations of conventional mass spectrometry. However,  the extremely high ionising efficiency of the ICP ion source of the MC-ICPMS now makes precise determination of the isotope ratios of these elements relatively straightforward.

We have recently developed procedures for in situ determination of copper isotope ratios in chalcopyrite using the new GEMOC LAM-MC-ICPMS system. An important new finding was that laser ablation/transport processes can cause very significant isotopic fractionation. This required development of operating conditions that minimise such fractionation and of a chalcopyrite standard to calibrate the analyses. The technique has been applied to a study of copper-bearing ores from a wide range of depositional settings, to establish the range of copper isotope fractionation in ore forming processes. A preliminary study on the Cadia porphyry copper-gold deposit examined how Cu isotope compositions vary on the scale of one ore system (figure 2).

Figure 1.  Epsilon 65Cu for chalcopyrite

This preliminary study has resulted in two major findings:

  • copper derived from igneous sources has a relatively limited spread of copper isotopic compositions (figure 1). However, notable 65Cu depletion occurs in low-temperature sedimentary deposits, suggesting that biogenic fractionation of copper may be an important process.
  • significant variations in copper isotopic compositions occur within the Cadia porphyry Cu-Au deposit (figure 2). Significantly, copper isotope variation correlates with alteration style and gold grade.
We are now applying similar techniques to other isotopic systems (e.g., Sb, Ag) and an application has been submitted to AMIRA for funding to study the isotopic systematics of copper and other important commodity metals (e.g., Fe, Zn, Ni., Mo). This work is expected to yield information that can be used as a tool to:
  • constrain the type of deposit that has contributed to a surface exploration anomaly
  • alert explorationists to the presence of different domains within a buried ore body
  • allow metals source(s) and fluid pathways in ore systems to be traced, thereby furthering our understanding of ore genesis.

Contact: Simon Jackson, Bill Griffin, Norm Pearson
Part of the Metallogenesis and & Technology Development programs. Based on the Honours Thesis study carried out by Andrew Botfield (1999).

Figure 2.  Variation of Cu isotope composition and gold grade with depth, Cadia deposit.

What is the origin of the diamonds from eastern Australia?

ECONOMIC CONCENTRATIONS of diamonds, both in situ and as alluvial deposits, are mainly associated with kimberlitic and lamproitic intrusions in Archean and Proterozoic continental crust worldwide. The kimberlites and lamproites have transported xenoliths and xenocrysts from the cool cratonic roots (>150 km depth) of the upper mantle to the upper crust so rapidly that the metamorphic minerals, including diamonds, do not have an opportunity to re-equilibrate to crustal temperatures and pressures.

Occurrences of diamond in metamorphic and ophiolitic rocks are of less commercial significance, but are important in understanding lithosphere processes. They provide evidence that rocks once at great depth have been tectonically transported to the Earth's surface. All of the tectonically emplaced metamorphic diamond deposits are younger than Proterozoic, and are characterised by micro-diamonds (< 1mm), or graphitised relics of larger diamonds. None contains unaltered macrodiamond.

Resorbed dodecahedral diamonds from Wellington, NSW.

In eastern Australia, diamonds occur in Cenozoic alluvium overlying Paleozoic and Mesozoic crust. Similar deposits are known from Burma, Thailand, Kalimantan, the Ural Mountains and California, but the origin of such diamond deposits is not well established. There is no trace of a primary source, nor are they associated with the usual diamond indicator minerals.

A wide variety of imaging and analytical techniques was used to characterise diamonds from a number of localities in New South Wales, and has distinguished two Groups (A and B) with distinctive characteristics and origins (Publication numbers 89, 110, 124). Group A diamonds are Precambrian (but represent two different episodes based on Re/Os age determinations (2.1Ma and 3.4 Ga, Pearson et al, 1998). Group A diamonds are similar to those from kimberlites and lamproites from cratonic regions worldwide. Group B diamonds have physical and chemical characteristics that are unique among diamonds worldwide and are inferred to have formed in a Paleozoic subduction slab. However, all diamonds show evidence of emplacement at the surface by igneous, rather than tectonic processes despite their inferred different origins of formation.

Group A diamonds are similar in their primary crystal forms, internal growth structures, mineral inclusion compositions and carbon isotope ratios to diamonds found in kimberlitic/lamproitic hosts in Archean and Proterozoic cratons worldwide. Mineral inclusions are dominantly peridotitic and carbon isotope ratios have a typical mantle range and distribution.

Furthermore, the Re-Os ages constrain the origin of these diamonds to formation in ancient mantle sources. This, along with the nature of the surface abrasion structures and radiation damage, suggests that the Group A diamonds may represent an older group of diamonds that have been in secondary collectors for a significant time. If this is so, it is feasible that the Group A diamonds may be derived from a number of primary sources (both in terms of the mantle and magmatic hosts).

A Group A diamond with strong plastic deformation

The age and location of the primary host rocks of the Group A diamonds has not been identified. There is evidence for possible Proterozoic lithospheric components in eastern Australia but none for Archean components. At this stage it is not possible to identify a source region or a time of transportation to the surface for the Group A diamonds. However, Veevers (1994) has traced Lower Permian fluvio-glacial drainage patterns from east Antarctica to southeastern Australia prior to the breakup of Gondwanaland, and has shown that the distribution of ca. 600 Ma zircons in the Lachlan Fold Belt sediments is consistent with their derivation from Antarctica. Therefore Precambrian Antarctic terrains are a possible location for diamond-bearing kimberlites or lamproites that could be the primary sources for the Group A diamonds.

A cathodoluminescence image of the internal structure of a Group B diamond showing concentric growth zoning, sector structures and strong deformation

Group B diamonds are unique among documented diamond suites worldwide in their combination of surface features, strained and irregular internal structures, enriched carbon and nitrogen isotopic signatures, and the eclogite/calc-silicate suite of mineral inclusions. These features are consistent with their formation in a subducting slab (Publications 89, 110, 124 and unpub. PhD Thesis by R. Davies). A provenance for the Group B diamonds could be linked to subduction and subsequent magmatic emplacement within eastern Australia related to Carboniferous to early Permian subduction and arc accretion in the New England Fold Belt. This is consistent with diamond emplacement ages obtained from mineral inclusions, which constrain transport of the diamonds to the Earth's surface between ~300 Ma and  ~220 Ma. Provided that the diamonds have a subduction origin, formation ages would be similar to emplacement ages because the stability of diamond in a subducting slab is transient, with only a small window in which they could survive.  The majority of Group B diamonds occurring on basement terrains of the New England Fold Belt (at Copeton and Bingara) and Group B diamonds at Wellington (on basement of the Lachlan Fold Belt) are less abundant and show stronger evidence of surface abrasion, possibly corresponding to more alluvial reworking. However, all Group B diamonds clearly have been transported. They show abrasion, are well sorted and occur in mature alluvium preserved by Cenozoic basalts.

A typical Group B diamond with an irregular resorbed form, corroded surfaces and a frosted pit.

Contact: Rondi Davies, Sue O'Reilly, Bill Griffin
Funded by: CRA (now Rio Tinto), GEMOC, ARC, Kennecott Canada
Part of GEMOC's Lithosphere Mapping and Metallogenesis strands.

Event signatures - tracing crustal growth with detrital zircons

1999 MARKED SOME BIG STEPS toward GEMOC's goal of understanding the links between mantle processes and crustal generation.  One was the development of a new approach to the analysis of crustal formation processes, using zircon from sediments.

Understanding the genesis of a block of crust means answering three big questions:

  • When was material added to the crust?
  • How did it get there - juvenile addition from the mantle,  or reworking of old crust?
  • What sort of material was added?
  • This knowledge traditionally requires a long and expensive effort - field mapping, petrography, geochemistry and isotopic dating.  In recent years, one shortcut has been to use ion-probe dating of zircon grains in sediments derived from the block of crust being studied.  The age spectra that come from this kind of study can answer the when? but not the how? or the what?

    GEMOC has developed a more integrated approach, based on the in situ analytical capabilities of the Geochemical Analysis Unit (see section on Technology Development).  Zircons are concentrated from a sediment sample, mounted and polished for  several stages of analysis.

    (1) GEMOC's Electron Microprobe is used to collect backscattered electron/ cathodoluminescence images of the internal structure of each grain (see cover), to recognise different growth zones.  At the same time, each grain is analysed for a series of trace elements (Hf, Y, U, Th).  With these elements, each grain can be assigned to a restricted range of rock types, using a series of discriminant plots (figure 1).

    (2) U-Pb ages for each grain (or zone) are measured using GEMOC's quadrupole ICPMS-laser ablation microprobe, to produce an age spectrum.  A rigorous analytical protocol provides U-Pb ages with precision and accuracy equivalent to ion-probe analysis, but at significantly lower cost and reduced analysis time.

    (3)  The Hf isotope composition of each grain is measured using GEMOC's new multi-collector ICPMS with laser-ablation microprobe (LAM-MC-ICPMS).  The Hf-isotope data have precision and accuracy comparable to that obtained using standard TIMS analysis on zircon concentrates.  These data, combined with the U-Pb ages, are used to identify the relative proportions of juvenile (mantle-derived) material and recycled older crust, and to place a minimum age on the recycled component.  As a by-product, the analysis gives concentrations of Yb and Lu, which help to identify the parent rock type for each grain.

    Figure 1.  Relation of trace element composition of zircon to host-rock chemistry

    The integration of these different types of data gives Event Signatures ˇ  detailed pictures of crustal formation events, including their age, the degree of crust-mantle interaction and crustal reworking, and the range of rock types generated at each stage. For example, data from Devonian sediments on E. Greenland (figure 2) , from a collaboration with Dr. T.-L. Knudsen, University of Oslo, show a major addition to the crust around 1900 Ma.  Other significant igneous events occurred around 1000 Ma and 500 Ma, but the Hf-isotope data indicate that these mainly involved reworking of the 1900 Ma crust.

    In addition to large-scale crustal-generation studies, applications include correlation between different terranes for mineral exploration and tectonic analysis, and mapping the provenance of basin sediments.  Studies of crustal genesis in the Yilgarn Craton and the Mt. Isa Province are planned for 2000, with support from industry.

    Contact:  Bill Griffin, Simon Jackson, Sue O'Reilly, Elena Belousova
    Funded by:  ARC, BHP Exploration
    Part of GEMOC's Crustal Generation, Metallogenesis and Geotectonics strands.

    Figure 2.  U-Pb age vs Hf isotope composition in zircons from East Greenland sandstones.

    What role did oceanic plateaux play in the Gondwana break-up?

    A NUMBER OF   OCEANIC PLATEAUX cover the southern Atlantic and southwest Indian Ocean region.  Their geographical positions can be reconstructed close to the zone of the Gondwana break-up between Antarctica, South America and Africa.  Were these plateaux parts of the pre-break-up continent or did they evolve within the newly formed oceanic crust?  As part of a collaborative project, we conducted an extensive seismic survey over one of the larger plateaux, the Agulhas Plateau south of South Africa, with the aim to help solve the questions about its crustal structure, origin, and role in a plate tectonic reconstruction context.  Seismic reflection data show clear indications of numerous volcanic extrusion centres with random distribution.  We are able to date this phase of voluminous volcanism to Late Cretaceous time, a period when numerous other Large Igneous Provinces (LIP) formed.  Travel time inversion of ocean-bottom seismograph records reveals a crust up to 25 km thick with velocities between 7.0 and 7.6 km/s for the lower 50-70% of its crustal column.  We do not find indications for continental affinity, as suggested in previous studies, but rather a predominantly oceanic origin of the Agulhas Plateau, similar to that inferred for the Northern Kerguelen and the Ontong-Java Plateaux.  In Late Cretaceous time, its main crustal growth was controlled by the proximity of the spreading centres and by the passage over the Bouvet hot spot at 80-100 m.y.  A new reconstruction of the southern Atlantic and southwest Indian Ocean region also demonstrates that neither the Agulhas Plateau nor the Maud Rise could have existed before the break-up.  Future work will include geophysical investigations of the Maud Rise, the Mozambique Plateau and the Crozet Plateau.

    Contact: Karsten Gohl
    Funded by: MURG in collaboration with the Alfred Wegener Institute, Germany.
    Part of GEMOC's Lithosphere Mapping, Crustal Generation and Geotectonics strands.

    Reconstruction of the South Atlantic and Southwest Indian Ocean region between (a) Early Cretaceous (chron M0) and (b) Late Cretaceous (chron 34).  The bold lines represent spreading centres at chrons M0 and 34.  A.P. and M.P. abbreviate the Agulhas and Mozambique Plateaux, respectively.  The Agulhas Plateau must have come into existence after the Falkland Plateau drifted from the Mozambique Ridge.

    Xenoliths from the Kerguelen islands: mantle metasomatism and continent formation

    THE KERGUELEN ARCHIPELAGO in the southern Indian Ocean sits atop a very large oceanic volcanic plateau, and the extensive basalts are intruded by granitic and syenitic rocks.  Is this a model for the initiation of continent formation?  An ongoing cooperation between GEMOC and the University of St. Etienne has concentrated on the study of xenoliths in the basaltic rocks, to map the structure and composition of the lithosphere beneath the plateau.

    The ultramafic and mafic xenoliths occur in the youngest and more alkaline lavas, and include many of the types of ultramafic and mafic xenoliths entrained by within-plate basalts in continental and oceanic settings. Type I spinel-bearing harzburgites and dunites are samples of the lithospheric mantle equilibrated in the spinel peridotite stability field beneath the thick oceanic crust (14-20 km). Three groups of ultramafic to mafic xenoliths are interpreted as deep crustal and upper mantle segregates from the basaltic magmas of the archipelago.  A clinopyroxene +orthopyroxene +spinel series varies from clinopyroxene-rich lherzolites, through pyroxenites to metagabbros ±garnet ± sapphirine. A second clinopyroxene+ilmenite+spinel series is represented by clinopyroxenites±garnet and by garnet+spinel+ilmenite metagabbros. A third clinopyroxene+ilmenite series is represented by garnet-bearing metagabbros.

    The Type I harzburgites can be divided into two distinct types: Cr-diopside-bearing protogranular harzburgites and Mg-augite-bearing poikilitic harzburgites. Both carry evidence for several mantle metasomatic events.  The first produced LREE-enrichment in both types of harzburgite and Cr-Na-rich Mg-augite ± phlogopite in the poikilitic ones, and is ascribed to metasomatic reactions between previously depleted harzburgites and highly alkaline magmas. The second type of metasomatism results in formation of alkali-rich (K2O: 1-10 wt%) feldspar + olivine (2) + Ti-chromite + Nb-rutile + ilmenite + Cr-armalcolite + Cr-Ca-armalcolite, in reaction zones close to opx and spinel or in thin veins cross-cutting the olivines. The formation of this exotic mineral assemblage is probably related to percolation of TiO2-rich, H2O-poor alkaline silicate melts into harzburgites. The dunites show similar mg#  and REE patterns to the poikilitic harzburgites. Dunites in some composite xenoliths are wall rocks to magmatic dykes of tholeiitic or alkaline affinity. The dunites appear to have formed by interaction between depleted harzburgites and basaltic melts, producing metasomatic reactions such as: (1) opx + liquid 1 ˇ> ol + liquid 2 and (2) opx + liquid 1 --> cpx + liquid 2.

    All the plagioclase-bearing xenoliths have reequilibrated under granulite facies conditions, in the Kerguelen oceanic lower crust and upper mantle. These mafic granulites are the first examples reported from an oceanic environment. The existence of oceanic granulites beneath the Kerguelen islands is consistent with the presence of a thickened crust detected by seismic studies, and the calculated and measured Vp values (collaboration with I. Jackson, RSES, ANU) for the basic granulites are similar to those observed in the low-velocity region beneath the oceanic crust (Vp = 7.2-7.5 k/s). Thus, this first study of the Kerguelen ultramafic and mafic xenoliths indicates that crustal thickening developed as basalts intruded at different levels of the lithosphere. The synergy between the young East-Indian Ridge and the Kerguelen hot spot (~ 45 Ma ago) produced voluminous tholeiitic-transitional magmas with resultant intrusion and formation of associated cumulates in the vicinity of the Moho. This process of crustal thickening by deep intrusion (underplating) was later amplified by the extrusive volcanic sequences related to the longevity of the hot spot activity (~ 45 Ma), leading to granulite-facies metamorphism of the lower crust.

    Publications:  117, 175 and 176
    Contact: Sue O'Reilly, Michel Grégoire
    Funded by: ARC, Macquarie University
    Part of GEMOC's Crustal Generation and Lithosphere Mapping strands.

    Model for the evolution of the Kerguelen lithosphere


    A demonstration of the power of the new laser ablation MC-ICPMS technology emerged late in 1999, as an unexpected byproduct of a study of Sm-Nd and Rb-Sr systems in apatite and titanite (sphene) from granitoid rocks in the Mt Isa inlier, done in collaboration with Dr. Geordie Mark of James Cook University.  The project is using trace element chemistry (by LAM-ICPMS) of apatite to link crystallisation and mineralisation processes in granitoids and associated hydrothermal systems.  The same samples were then analysed for Sr and Nd isotopes using the LAM-MC-ICPMS.  To obtain the initial Nd isotope composition of the granitoids, we analysed separated grains of titanite; the analysis gives both the Sm-Nd ratio and the isotopic composition of the Nd.  At the levels of Nd in these grains (2000-5000 ppm) the in situ analysis of single points typically produces data with precision and accuracy equivalent to conventional TIMS analysis of bulk samples (143Nd/144Nd ± 0.00003 (2sd)).  In the titanite concentrate from the Mt. Angelay quartz monzonite, we found an unexpectedly large degree of heterogeneity in both Sm/Nd and Nd isotopic composition. The data define an excellent isochron (MSWD = 1.7)  with an age of 1533 ±70 (2sd), identical to a SHRIMP zircon U-Pb age on the quartz monzonite of 1525 Ma.  The initial 143Nd/144Nd of 0. 51030±6 is similar to the value determined by whole-rock analysis of samples from the same intrusion.  The TDM model age of ca 2500 Ma indicates a late Archean age for the source of the granitoid.  The precision of the isochron age and the initial ratio, despite the relatively small spread in Sm/Nd, attests to the accuracy of the individual analyses.

    The total analysis time used for this work was two hours.  The technique may be applicable to a wider range of samples, and could offer a rapid and cost-effective tool for reconnaissance geochronological studies, and for determining the age and provenance of detrital titanites in sediments.

    Contact:  Bill Griffin, Geordie Mark (JCU)
    Funded by:  ARC and AMIRA
    Part of GEMOC's Crustal Generation strand.

    Shell Games:  Sr isotopes and Permian Terranes in New Zealand

    THE ISOTOPIC COMPOSITIONof Sr in marine fossils reflects that of the sea water in which they grew, and the large variations in the 87Sr/86Sr of sea water throughout the Phanerozoic offer an indirect method of dating calcareous sediments. Previously, this approach has wrestled with problems of identifying pristine carbonate material.  The analysis of the Sr also commonly required laborious and time-consuming microdrilling of individual fossil fragments, followed by chemical separation of Sr and analysis by conventional mass-spectrometry. The new LAM-MC-ICPMS technique overcomes many of these problems, by providing rapid high-precision in situ Sr isotope analyses on a 200-micron scale, so that isotopic variations within single fossil fragments, and differences or similarities between fossils and matrix, can be used to recognise fossils whose isotopic composition has been modified by diagenesis.

    Provenance studies of the Permian-Jurassic volcanic arc terranes of New Zealand require dating of the extensive volcaniclastic sediments, many of which contain few fossils.  However, the distinctive prismatic calcite shell debris of the bivalve Atomodesma is widespread from Early-Late Permian, and over this period the 87Sr/86Sr of sea water fell from 0.7081 (275 Ma) to 0.7068 (265 Ma), then rose to ca 0.7071 (251 Ma).  LAM-MC-ICPMS analyses of Sr isotopes were done on Atomodesma debris from several known sections, and a variety of matrix types, to test the methods and the stratigraphic usefulness of the data.

    Atomodesma fragments contained 300-1200 ppm Sr.  A laser beam of 200-400 microns in diameter, rastered over a 1-5 mm length, gave signals high enough to provide 87Sr/86Sr data with an external precision better than ±0.00004 (2sd).  This is equivalent to a stratigraphic precision of ca  ±300,000 years, which is very useful.  At a given stratigraphic time horizon, we found consistent 87Sr/86Sr ratios for Atomodesma fragments, irrespective of their matrices; in one case concordant results were found for coexisting echinoid debris.  Large differences were found between the fossils and their clastic matrices.  The data obtained so far are consistent with the accepted sea-water curve for the Permian, and show that Atomodesma, while not a very good marker fossil in itself, can be used for detailed stratigraphic correlation by analysis of its 87Sr/86Sr

    Contact:  Bill Griffin, Dr. C.J. Adams (Inst. Geological & Nuclear Sciences, Lower Hutt, NZ (, Norm Pearson
    Funded by: GEMOC and the Institute of Geological and Nuclear Sciences
    Part of GEMOC's Crustal Genesis strand.

    Section of Tramway Sandstone containing marine fossils (the bivalve Atomodesma, marked in black).  The whitish grooves within the fossils are laser ablation tracks produced during Sr isotope analysis using GEMOC's LAM-MC-ICPMS. Width of photo = 1.5 cm

    Defining the APWP for Paleozoic eastern Gondwana paleopoles from the Northen TASMAN orogen, Queensland

    THE MID- CARBONIFEROUS segment of Gondwana's Apparent Polar Wander Curve (APWP) is still poorly defined and it is suggested that resolution of the preliminary data presented herein could provide a key pole in the definition of this part of the path.  Ignimbritic sequences from the Permo-Carboniferous Newcastle Range (NRV), central Queensland (18.3 S, 143.7 E), were sampled for paleomagnetic investigation.  Mid-Carboniferous volcanics display a characteristic magnetisation direction for the majority of sites: DC=202.7, IC=62.6, with an antipodal direction of DC=8.7, IC=-57.5.  These directions indicate a mid-Carboniferous paleomagnetic pole of Plat=-58.6, Plong=115.8.  This well-defined pole is supported by results from Africa and South America as reported in the Global Paleomagnetic Database. (See figure)

    Sediments from a site in the Devonian Gilberton Formation (DGF) exhibited two magnetic components, one which is described as the mid-Carboniferous overprint (DC=191.1, IC=55.7) and the second, a hematite (high-temperature) component, has a direction of DD=355.9, ID=-10.  This high-temperature component gave a VGP of Plat=-76.1, Plong=306.5, a pole that is, seemingly, in concordance with that of the Late Devonian Comerong Volcanics (CV), Plat=-76.9, Plong=330.7.

    The coupling of paleomagnetic analysis with detailed rock magnetic tests is an integral part of the geophysics research conducted at Macquarie.  An enhanced understanding of paleogeodynamics is the aim of this program for which the collaboration between GEMOC and CSIRO is vital.  Mid-2000 will see an additional field excursion to northern Queensland, followed by one to the Georgina Basin of far western Queensland.  It is this second trip that will best exemplify the collaborative effort that is necessary in paleogeodynamics as reinvestigation of the Cambrian/Ordovician Black Mountain carbonates, key units in the recently proposed IITPW event, will depend heavily upon chemical analyses for correlation of paleomagnetisations with earlier studies.

    Contact:  Kari Anderson Mark Lackie
    Funded by:  Macquarie University Small ARC
    Part of GEMOC's Geotectonics strand.

    Late Paleozoic APWP for Australia

    Reading Granite Recipes from Zircon Tapes

    ISOTOPIC ANALYSIS of  elements such as Sr, Nd and Hf is widely used to evaluate the sources of igneous rocks.  Many granites have isotopic compositions that imply they are made up of material from two or more sources, such as remelted older rocks and juvenile mantle material (see Event Signatures).  Unfortunately, these analyses only tell us the final state, and leave us in the dark about when and how these different components got together to make up the magma that crystallised into the sample .

    But many of these rocks contain grains of zircon, and the zoning in single zircon crystals acts like a tape recorder, preserving a record of changes in magma composition and physical conditions such as temperature.  The challenge has been to read this information at a useful scale.

    Figure 1.  BSE/CL image of a zircon, showing changes of crystal morphology during growth

    Backscattered electron/cathodoluminescence images (figure 1) of sectioned zircon grains show changes in zircon growth forms that reflect changes in the crystallisation environment, and detailed work by GEMOC, using electron microprobe and LAM-ICPMS techniques, has documented large changes in trace-element patterns from zone to zone.  These changes appear to reflect major shifts in magma composition, which can be related to mixing of different magmas, or to changes in fractional crystallisation patterns.

    GEMOC's new multi-collector ICPMS, with its attached laser microprobe, lets us test whether magma mixing was being recorded in such zircons. Crystallisation of a single magma can produce changes in magma composition, but will not affect the isotopic composition of a single element such as Hf, which is abundant in zircon.  However, mixing of two magmas with different sources could produce a change in the isotopic composition of Hf.  Isotopic heterogeneity in the Hf of the zircons from one rock would be good evidence of magma mixing.

    As part of GEMOC's China collaboration program, we have carried out a detailed morphological and geochemical study of zircons from a series of rocks from the Jurassic Pingtan igneous complex in Fujian, China, where field relations indicate the mingling of mafic and felsic magmas.  When the LAM-MC-ICPMS was used to analyse Hf isotopes in these zircons, we found significant heterogeneity from grain to grain, and from zone to zone within single grains (figure 2).  The data show that relatively felsic magmas, derived from remelting of crustal rocks  600-800 Ma older, were mixed with more juvenile magmas derived from the depleted mantle, during their crystallisation.  Such mixing may have occurred several times.

    These are the first studies of this kind.  This new type of data provides a new dimension to studies of magma genesis and mechanisms of crust-mantle interaction, and illustrates how the new LAM-MC-ICPMS technology is going to revolutionise many areas of petrology and geochemistry in the next few years.

    Contact:  Bill Griffin, Simon Jackson, Wang Xiang and Sue O'Reilly.  Part of GEMOC's Crustal Generation strands.  Funded by ARC and AusAID.

    Figure 2.  Cumulative probability curve of Hf isotope compositions in zircons from a Pingtan quartz diorite.  Arrows show zoning within single grains.